Collections as Data Facets document collections as data implementations. An implementation consists of the people, services, practices, technologies, and infrastructure that aim to encourage computational use of cultural heritage collections.
Mary Elings, University of California Berkeley, Bancroft Library
Quinn Dombrowski, University of California Berkeley, Research IT
1. Why do it
an accessible narrative description that describes why the decision was made to do the work
In April 2014 to celebrate the 50th anniversary of the Free Speech Movement at UC Berkeley, The Bancroft Library, the Research IT group in the Office of the CIO, and the School of Information at UC Berkeley held #HackFSM, a hackathon around the Free Speech Movement Digital Archive, as part of the Digital Humanities @ Berkeley initiative. The event brought together thirteen teams of UC Berkeley students to design a new interface for a subset of Bancroft’s digital holdings on the Free Speech Movement.
The Free Speech Movement was an appealing, immediately recognizable subject of the hackathon. The Free Speech Movement is felt to be quintessentially “Berkeley”, and while most students are aware of the movement, it is not necessarily well understood by those students. The hackathon offered an opportunity to raise awareness of the subject and there was an available dataset to work with in the Bancroft Library’s Free Speech Movement (FSM) digital archive.
2. Making the Case
an accessible narrative description that describes how the administrative case was made to do the work
The hackathon served as a valuable opportunity for groups in very different areas of the university, with different priorities and organizational cultures, to work together towards a shared vision. There were areas of administrative overlap, particularly between the Library and Research IT groups, and clearly defining roles and responsibilities was essential. #HackFSM was a highly collaborative and interdisciplinary effort, made possible by the participation of the Library Systems Office, Library Administration, BIDS, the School of Information, Arts & Humanities Division, Social Sciences, and the students from various disciplines, in addition to the Bancroft Library and Research IT. The relationships formed through participating in this hackathon have continued to benefit campus through the development of new collaborative initiatives.
3. How you did it
an accessible narrative description that describes how you did it; who was involved - what their roles were; what services were drawn upon; what collections were involved and why were they selected; what infrastructure and technologies were selected and why; what challenges were encountered in the course of the implementation
see the white paper (below)
4. Share the docs
an assortment of formal documentation (personas, use cases, functional requirements), workflows, and code you have that supports the implementation
This white paper describes the process of organizing #HackFSM, a digital humanities hackathon around the Free Speech Movement digital archive, jointly organized by Research IT at UC Berkeley and The Bancroft Library. The paper includes numerous appendices and templates of use for organizations that wish to hold a similar event.
5. Understanding use
discrete discussion of how you approach understanding use (e.g. library use, research use, pedagogical use, creative use, etc.) - discussion of exemplary use is also encouraged
There was never an explicit discussion of “use”; it was left up to the individual student teams to define the audience for their project, and what “use” looked like. Responses varied, and included a tool for conducting research, multiple browsing / exploration interfaces, and a few that were more like an exhibit.
6. Who supports use
discrete discussion of the people, services, and programs that support use of the data (e.g. digital scholarship services, instruction services, subject area liaisons, etc.)
The HackFSM team included The Bancroft Library, the Research IT group in the Office of the CIO, and the School of Information at UC Berkeley. The data preparation for the API involved the Library Systems Office and the Bancroft Library. In order to govern access to the Library’s FSM API, ResearchIT staff used a common-good campus service (no cost to users) called API Central, provided by UC Berkeley’s Information Services and Technology department. The API Central service provides a proxy to the Solr API, and can be configured to require credentials in order to process an HTTP Request (credentials are values of app_id and app_key headers that are set in the HTTP Request Header). University IT staff, I-School faculty, Berkeley alumni, and individuals from local tech companies served as code mentors during the hackathon. Eventbrite was used for registration of participants. Social media accounts (twitter and Facebook) were used to promote the event. During the hacking period, students, mentors, and event organizers communicated via Piazza, a free platform that offers a course-based message board, commonly used in STEM courses at UC Berkeley.
The Library administration offered space, as the new Berkeley Institute for Data Science space and the UC Berkeley School of Information for the opening and closing events. During the hackathon students were encouraged to make use of physical collaboration space provided by our new social sciences D-Lab and library.
7. Things people should know
distilled things that people should know if they are thinking about pursuing similar work
Projects like this are highly collaborative and require technologists as well as content providers. The most successful outcome of the project was student engagement. Students from across disciplines came together to build something.
Maintaining the winning sites was not successful and we need better method and practices to achieve a record of this work.
While the main work product was a website, the greater product was that developers and humanists learned to communicate and work together. IT was humanists and technologists working and talking together, learning from and collaborating with each other in the process of building new scholarly output. Hopefully events like HackFSM can prepare them for future collaborations in a research environment where such interdisciplinary projects will be more common.
8. Whats next
you have something in place. whats next and why?
Our hope is to prepare more digitized collections as data so they are ready to be used computationally. Current OCR could be improved and brought to a point of being “research ready” for computational use. We plan to write a grant to prepare a large recently digitized archival collection, working with local data scientists on the requisite steps we would need to take to get the data to a point of usefulness.